390 SEXUAL REPRODUCTION 



secondly as a possible means of introducing new characters or 

 of rearranging those already present in the two parents. The 

 former result may be attained by other stimuh. Amongst 

 animals, for instance, the eggs of the Frog have been caused 

 to develop bj^ mere pricking with a glass needle covered with 

 blood-serum, and those of Sea-urchins by treatment with solu- 

 tions having a higher osmotic pressure than sea-water. Amongst 

 plants the polyembryonic seeds of the Orange (cf. p. 372) furnish 

 an analogous example ; here certain cells, apart from the ferti- 

 lised egg, have developed into embryos, but there is no evidence 

 that the apogamously produced seedlings are any less vigorous 

 than those resulting from sexual fusion. It seems probable, 

 therefore, that the chief advantage of sexual reproduction lies 

 in the possibility of producing organisms, with a slightly different 

 hereditary constitution, such as may survive under conditions 

 that would be unfavourable to the pure parent strain. In 

 other words, sexual reproduction provides material upon which 

 natural selection can operate. 



The segregation of characters above referred to is most prob- 

 ably effected during the reduction division in the spore mother- 

 cells. In this process the chromosomes, instead of splitting 

 longitudinally, as in the vegetative divisions, separate in their 

 entirety into two sets. These pass to the respective daughter- 

 nuclei, so that, of the four resulting spores, two possess the 

 characters borne by one set, and two those borne by the other 

 set. If this hypothesis be true, the two separate sets of chromo- 

 somes probably represent the paternal and maternal chromatin 

 respectively. In the case of an allelomorphic pair, it is assumed 

 that the dominant character is present in one set and the re- 

 cessive in the other. In the reduction division, therefore, the 

 allelomorphs will become separated, so that pure recessives and 

 pure dominants can be bred. Such a theoretical conception is 

 incapable of proof, and it is only warranted because it tallies 

 with the observed facts. 



In vegetative propagation the offspring normall}^ exhibit no 

 change of character, as compared with the parent, and new 

 forms can only arise bj' mutation. Mutations in vegetatively 

 produced offspring, and even in certain branches of an individual, 

 have indeed been occasionally observed (cf. p. 3S2), and may 



